Kari Stefansson’s research while affiliated with Reykjavík University and other places

Background Abnormal plasma sodium concentration represents an imbalance of total body water relative to electrolyte content. Hyponatremia is a common and potentially severe adverse event, and thiazide diuretics constitute a leading cause of drug-induced hyponatremia. Methods We conducted genome-wide association study analyses of plasma sodium concentration, thiazide - induced decrease in sodium concentration, and thiazide - induced hyponatremia in a total of 188,464 individuals of European ancestry . Additionally, we tested for gene-environment interaction between a polygenic score developed for plasma sodium concentration and thiazide exposure on sodium concentration and hyponatremia risk. Results Meta-analysis yielded 31 independent associated signals at P <5×10 ⁻⁸ with plasma sodium concentrations. Subsequent tissue specificity analysis showed a significantly increased expression of sodium-associated genes in pituitary tissue ( P =4.5×10 ⁻⁵ ). No genome-wide significant loci were found for thiazide - induced sodium concentration decrease or thiazide-induced hyponatremia. A polygenic score for plasma sodium concentration was associated with 0.43 (95% confidence interval (CI) = 0.39-0.46) mmol/L lower plasma sodium per standard deviation lower, and thiazide use was associated with 0.80 (95% CI=0.72-0.88) mmol/L lower plasma sodium, but we observed no gene-environment interaction effect ( P =0.71). Conclusions These results underline the role of genetic variation in regulating plasma sodium concentration and highlight the importance of pathways involving the pituitary gland while finding no evidence of genetic predisposition for the plasma sodium-lowering effect of thiazides.

Human recombination maps are a valuable resource for association and linkage studies and crucial for many inferences of population history and natural selection. Existing maps1, 2, 3, 4–5 are based solely on cross-over (CO) recombination, omitting non-cross-overs (NCOs)—the more common form of recombination⁶—owing to the difficulty in detecting them. Using whole-genome sequence data in families, we estimate the number of NCOs transmitted from parent to offspring and derive complete, sex-specific recombination maps including both NCOs and COs. Mothers have fewer but longer NCOs than fathers, and oocytes accumulate NCOs in a non-regulated fashion with maternal age. Recombination, primarily NCO, is responsible for 1.8% (95% confidence interval: 1.3–2.3) and 11.3% (95% confidence interval: 9.0–13.6) of paternal and maternal de novo mutations, respectively, and may drive the increase in de novo mutations with maternal age. NCOs are substantially more prominent than COs in centromeres, possibly to avoid large-scale genomic changes that may cause aneuploidy. Our results demonstrate that NCOs highlight to a much greater extent than COs the differences in the meiotic process between the sexes, in which maternal NCOs may reflect the safeguarding of oocytes from infancy until ovulation.

Investigating the genetic factors influencing human birth weight may lead to biological insights into fetal growth and long-term health. We report analyses of rare variants that impact birth weight when carried by either fetus or mother, using whole exome sequencing data in up to 234,675 participants. Rare protein-truncating and deleterious missense variants are collapsed to perform gene burden tests. We identify 9 genes; 5 with fetal-only effects on birth weight, 1 with maternal-only effects, 3 with both, and observe directionally concordant associations in an independent sample. Four of the genes were previously implicated by GWAS of birth weight. IGF1R and PAPPA2 (fetal and maternal-acting) have known roles in insulin-like growth factor bioavailability and signalling. PPARG, INHBE and ACVR1C (fetal-acting) are involved in adipose tissue regulation, and the latter two also show associations with favourable adiposity patterns in adults. We highlight the dual role of PPARG (fetal-acting) in adipocyte differentiation and placental angiogenesis. NOS3 (fetal and maternal-acting), NRK (fetal), and ADAMTS8 (maternal-acting) have been implicated in placental function and hypertension. To conclude, our analysis of rare coding variants identifies regulators of fetal adipose tissue and fetoplacental angiogenesis as determinants of birth weight, and further evidence for the role of insulin-like growth factors.

Iron homoeostasis is tightly regulated, with hepcidin and soluble transferrin receptor (sTfR) playing significant roles. However, the genetic determinants of these traits and the biomedical consequences of iron homoeostasis variation are unclear. In a meta-analysis of 12 cohorts involving 91,675 participants, we found 43 genomic loci associated with either hepcidin or sTfR concentration, of which 15 previously unreported. Mapping to putative genes indicated involvement in iron-trait expression, erythropoiesis, immune response and cellular trafficking. Mendelian randomisation of 292 disease outcomes in 1,492,717 participants revealed associations of iron-related loci and iron status with selected health outcomes across multiple domains. These associations were largely driven by HFE, which was associated with the largest iron variation. Our findings enhance understanding of iron homoeostasis and its biomedical consequences, suggesting that lifelong exposure to higher iron levels is likely associated with lower risk of anaemia-related disorders and higher risk of genitourinary, musculoskeletal, infectious and neoplastic diseases.

Mutations that affect RNA splicing significantly impact human diversity and disease. Here we present a method using transformers, a type of machine learning model, to detect splicing from raw 45,000-nucleotide sequences. We generate embeddings with residual neural networks and apply hard attention to select splice site candidates, enabling efficient training on long sequences. Our method surpasses the leading tool, SpliceAI, in detecting splice sites in GENCODE and ENSEMBL annotations. Using extensive RNA sequencing data from an Icelandic cohort of 17,848 individuals and the Genotype-Tissue Expression (GTEx) project, our method demonstrates superior performance in detecting splice junctions compared to SpliceAI-10k (PR-AUC = 0.834 vs. PR-AUC = 0.820) and is more effective at identifying disease-related splice variants in ClinVar (PR-AUC = 0.997 vs. PR-AUC = 0.996). These advancements hold promise for improving genetic research and clinical diagnostics, potentially leading to better understanding and treatment of splicing-related diseases.

Dilated cardiomyopathy (DCM) is a leading cause of heart failure and cardiac transplantation. We report a genome-wide association study and multi-trait analysis of DCM (14,256 cases) and three left ventricular traits (36,203 UK Biobank participants). We identified 80 genomic risk loci and prioritized 62 putative effector genes, including several with rare variant DCM associations ( MAP3K7 , NEDD4L and SSPN ). Using single-nucleus transcriptomics, we identify cellular states, biological pathways, and intracellular communications that drive pathogenesis. We demonstrate that polygenic scores predict DCM in the general population and modify penetrance in carriers of rare DCM variants. Our findings may inform the design of genetic testing strategies that incorporate polygenic background. They also provide insights into the molecular etiology of DCM that may facilitate the development of targeted therapeutics.

Myasthenia gravis (MG) is a rare autoantibody-mediated disease affecting the neuromuscular junction. We performed a genome-wide association study of 5708 MG cases and 432,028 controls of European ancestry and a replication study in 3989 cases and 226,643 controls provided by 23andMe Inc. We identified 12 independent genome-wide significant hits (P < 5e⁻⁸) across 11 loci. Subgroup analyses revealed two of these were associated with early-onset (at age <50) and four with late-onset MG (at age ≥ 50). Imputation of human leukocyte antigen alleles revealed inverse effect sizes for late- and early-onset, suggesting a potential modulatory influence on the time of disease manifestation. We assessed the performance of polygenic risk scores for MG, which significantly predicted disease status in an independent target cohort, explaining 4.21% of the phenotypic variation (P = 5.12e⁻⁹). With this work, we aim to enhance our understanding of the genetic architecture of MG.